Is it possible to customize the configuration for the speech recognition model?

The background is that I used Qianwen ASR with native support, but after entering the API Key, I received a billing statement from Qianwen that same evening.However, to my knowledge, Qianwen itself offers free model trial quotas. These require configuring model versions with date-based identifiers like 0910. Since our current setup doesn't support flexible configuration, we cannot utilize the free quota and must pay directly. This user experience design seems somewhat unreasonable. Additionally, if support for multiple ASR models were available, we could also try integrating other vendors' APIs for comparative testing.